Topics

Don't fix if it ain't broke or preventative maintenance

Dwayne Verhey <yg@...>
 

I joined this group for advice and insight for a newly-acquired 475 of unknown provenance, and with your help, I found most of the problems were with me, not the scope; only one minor repair, cleaning and some tweaking were actually required to get it up and running.

As I lurk, about 99% of the repair advice here goes straight over my head but I note a couple of recurring themes. Tektronix scopes may contain custom, irreplaceable, components and failure of one component may take out other bits and pieces.

As with other old electronics, capacitors seem to be mentioned a lot in that second category. Particularly electrolytic and tantalum types. It's almost automatic to replace original electrolytic capacitors in most other old electronics, but I've seen two schools of thought with regards to Tek scopes. A recent comment here suggested "don't fix it if it ain't broke", and I have read that Tek used top quality components coupled with circuit designs that did not stress the components, so failures are supposedly the exception rather than the rule. And there's always the risk of a ham-fisted amateur (ie: me) damaging things like traces when replacing components. But then, there's a lot of people talking about dry electrolytics and the smell of tantalum around here...

So what are your thoughts: would I be better off attempting some preventative maintenance like replacing all the electrolytics or leaving well enough alone until/unless something actually breaks?

<cross fingers and hope this doesn't hijack an existing thread>

Dwayne

 

For aluminum electrolytic capacitors, I leave them alone until one or
more fails or there are symptoms indicating that they are worn and
then I change all of them starting with the output capacitors for
switching power supplies and the input capacitors for linear power
supplies.

If they are old, I may change them during the process of repairing
something else while I have access anyway.

Solid tantalum capacitors are much the same except that they do not
have a real wear out mechanism and seem to fail destructively at which
point I will change any of the same type which were not properly
voltage derated.

On Fri, 01 Jan 2016 21:27:58 -0500, you wrote:

I joined this group for advice and insight for a newly-acquired 475 of
unknown provenance, and with your help, I found most of the problems
were with me, not the scope; only one minor repair, cleaning and some
tweaking were actually required to get it up and running.

As I lurk, about 99% of the repair advice here goes straight over my
head but I note a couple of recurring themes. Tektronix scopes may
contain custom, irreplaceable, components and failure of one component
may take out other bits and pieces.

As with other old electronics, capacitors seem to be mentioned a lot in
that second category. Particularly electrolytic and tantalum types. It's
almost automatic to replace original electrolytic capacitors in most
other old electronics, but I've seen two schools of thought with regards
to Tek scopes. A recent comment here suggested "don't fix it if it ain't
broke", and I have read that Tek used top quality components coupled
with circuit designs that did not stress the components, so failures are
supposedly the exception rather than the rule. And there's always the
risk of a ham-fisted amateur (ie: me) damaging things like traces when
replacing components. But then, there's a lot of people talking about
dry electrolytics and the smell of tantalum around here...

So what are your thoughts: would I be better off attempting some
preventative maintenance like replacing all the electrolytics or leaving
well enough alone until/unless something actually breaks?

<cross fingers and hope this doesn't hijack an existing thread>

Dwayne

petertech99h
 

Dwayne,
I'd go 50/50, inspect caps for corrosion, bulging cases, etc and start sourcing replacement parts as you go and set them aside so they'll be there when you need them.  Some caps on double sided PCBs have plate thru holes where the solder is hard to suck out without damage to the board, so upgrade your solder iron and solder removal tools. Make sure you have the right solder for your Tek scope, the very old vintage scopes have silver bearing solder! If there's a lead-free solder requirement, practice since this style solder is not so nice to work with as 63/37 lead solders.

In my case, I check the AC ripple voltage across the caps on my scope's linear power supplies, when it gets higher then it should I know a cap is starting to fail.On 'newer' Tek scopes with switch-mode power supplies you can't really do that though.
Have fun!
Pete

On Friday, January 1, 2016 10:08 PM, "David @DWH [TekScopes]" <TekScopes@...> wrote:


  For aluminum electrolytic capacitors, I leave them alone until one or
more fails or there are symptoms indicating that they are worn and
then I change all of them starting with the output capacitors for
switching power supplies and the input capacitors for linear power
supplies.

If they are old, I may change them during the process of repairing
something else while I have access anyway.

Solid tantalum capacitors are much the same except that they do not
have a real wear out mechanism and seem to fail destructively at which
point I will change any of the same type which were not properly
voltage derated.

On Fri, 01 Jan 2016 21:27:58 -0500, you wrote:

I joined this group for advice and insight for a newly-acquired 475 of
unknown provenance, and with your help, I found most of the problems
were with me, not the scope; only one minor repair, cleaning and some
tweaking were actually required to get it up and running.

As I lurk, about 99% of the repair advice here goes straight over my
head but I note a couple of recurring themes. Tektronix scopes may
contain custom, irreplaceable, components and failure of one component
may take out other bits and pieces.

As with other old electronics, capacitors seem to be mentioned a lot in
that second category. Particularly electrolytic and tantalum types. It's
almost automatic to replace original electrolytic capacitors in most
other old electronics, but I've seen two schools of thought with regards
to Tek scopes. A recent comment here suggested "don't fix it if it ain't
broke", and I have read that Tek used top quality components coupled
with circuit designs that did not stress the components, so failures are
supposedly the exception rather than the rule. And there's always the
risk of a ham-fisted amateur (ie: me) damaging things like traces when
replacing components. But then, there's a lot of people talking about
dry electrolytics and the smell of tantalum around here...

So what are your thoughts: would I be better off attempting some
preventative maintenance like replacing all the electrolytics or leaving
well enough alone until/unless something actually breaks?

<cross fingers and hope this doesn't hijack an existing thread>

Dwayne
#yiv3297290559 #yiv3297290559 -- #yiv3297290559ygrp-mkp {border:1px solid #d8d8d8;font-family:Arial;margin:10px 0;padding:0 10px;}#yiv3297290559 #yiv3297290559ygrp-mkp hr {border:1px solid #d8d8d8;}#yiv3297290559 #yiv3297290559ygrp-mkp #yiv3297290559hd {color:#628c2a;font-size:85%;font-weight:700;line-height:122%;margin:10px 0;}#yiv3297290559 #yiv3297290559ygrp-mkp #yiv3297290559ads {margin-bottom:10px;}#yiv3297290559 #yiv3297290559ygrp-mkp .yiv3297290559ad {padding:0 0;}#yiv3297290559 #yiv3297290559ygrp-mkp .yiv3297290559ad p {margin:0;}#yiv3297290559 #yiv3297290559ygrp-mkp .yiv3297290559ad a {color:#0000ff;text-decoration:none;}#yiv3297290559 #yiv3297290559ygrp-sponsor #yiv3297290559ygrp-lc {font-family:Arial;}#yiv3297290559 #yiv3297290559ygrp-sponsor #yiv3297290559ygrp-lc #yiv3297290559hd {margin:10px 0px;font-weight:700;font-size:78%;line-height:122%;}#yiv3297290559 #yiv3297290559ygrp-sponsor #yiv3297290559ygrp-lc .yiv3297290559ad {margin-bottom:10px;padding:0 0;}#yiv3297290559 #yiv3297290559actions {font-family:Verdana;font-size:11px;padding:10px 0;}#yiv3297290559 #yiv3297290559activity {background-color:#e0ecee;float:left;font-family:Verdana;font-size:10px;padding:10px;}#yiv3297290559 #yiv3297290559activity span {font-weight:700;}#yiv3297290559 #yiv3297290559activity span:first-child {text-transform:uppercase;}#yiv3297290559 #yiv3297290559activity span a {color:#5085b6;text-decoration:none;}#yiv3297290559 #yiv3297290559activity span span {color:#ff7900;}#yiv3297290559 #yiv3297290559activity span .yiv3297290559underline {text-decoration:underline;}#yiv3297290559 .yiv3297290559attach {clear:both;display:table;font-family:Arial;font-size:12px;padding:10px 0;width:400px;}#yiv3297290559 .yiv3297290559attach div a {text-decoration:none;}#yiv3297290559 .yiv3297290559attach img {border:none;padding-right:5px;}#yiv3297290559 .yiv3297290559attach label {display:block;margin-bottom:5px;}#yiv3297290559 .yiv3297290559attach label a {text-decoration:none;}#yiv3297290559 blockquote {margin:0 0 0 4px;}#yiv3297290559 .yiv3297290559bold {font-family:Arial;font-size:13px;font-weight:700;}#yiv3297290559 .yiv3297290559bold a {text-decoration:none;}#yiv3297290559 dd.yiv3297290559last p a {font-family:Verdana;font-weight:700;}#yiv3297290559 dd.yiv3297290559last p span {margin-right:10px;font-family:Verdana;font-weight:700;}#yiv3297290559 dd.yiv3297290559last p span.yiv3297290559yshortcuts {margin-right:0;}#yiv3297290559 div.yiv3297290559attach-table div div a {text-decoration:none;}#yiv3297290559 div.yiv3297290559attach-table {width:400px;}#yiv3297290559 div.yiv3297290559file-title a, #yiv3297290559 div.yiv3297290559file-title a:active, #yiv3297290559 div.yiv3297290559file-title a:hover, #yiv3297290559 div.yiv3297290559file-title a:visited {text-decoration:none;}#yiv3297290559 div.yiv3297290559photo-title a, #yiv3297290559 div.yiv3297290559photo-title a:active, #yiv3297290559 div.yiv3297290559photo-title a:hover, #yiv3297290559 div.yiv3297290559photo-title a:visited {text-decoration:none;}#yiv3297290559 div#yiv3297290559ygrp-mlmsg #yiv3297290559ygrp-msg p a span.yiv3297290559yshortcuts {font-family:Verdana;font-size:10px;font-weight:normal;}#yiv3297290559 .yiv3297290559green {color:#628c2a;}#yiv3297290559 .yiv3297290559MsoNormal {margin:0 0 0 0;}#yiv3297290559 o {font-size:0;}#yiv3297290559 #yiv3297290559photos div {float:left;width:72px;}#yiv3297290559 #yiv3297290559photos div div {border:1px solid #666666;height:62px;overflow:hidden;width:62px;}#yiv3297290559 #yiv3297290559photos div label {color:#666666;font-size:10px;overflow:hidden;text-align:center;white-space:nowrap;width:64px;}#yiv3297290559 #yiv3297290559reco-category {font-size:77%;}#yiv3297290559 #yiv3297290559reco-desc {font-size:77%;}#yiv3297290559 .yiv3297290559replbq {margin:4px;}#yiv3297290559 #yiv3297290559ygrp-actbar div a:first-child {margin-right:2px;padding-right:5px;}#yiv3297290559 #yiv3297290559ygrp-mlmsg {font-size:13px;font-family:Arial, helvetica, clean, sans-serif;}#yiv3297290559 #yiv3297290559ygrp-mlmsg table {font-size:inherit;font:100%;}#yiv3297290559 #yiv3297290559ygrp-mlmsg select, #yiv3297290559 input, #yiv3297290559 textarea {font:99% Arial, Helvetica, clean, sans-serif;}#yiv3297290559 #yiv3297290559ygrp-mlmsg pre, #yiv3297290559 code {font:115% monospace;}#yiv3297290559 #yiv3297290559ygrp-mlmsg * {line-height:1.22em;}#yiv3297290559 #yiv3297290559ygrp-mlmsg #yiv3297290559logo {padding-bottom:10px;}#yiv3297290559 #yiv3297290559ygrp-msg p a {font-family:Verdana;}#yiv3297290559 #yiv3297290559ygrp-msg p#yiv3297290559attach-count span {color:#1E66AE;font-weight:700;}#yiv3297290559 #yiv3297290559ygrp-reco #yiv3297290559reco-head {color:#ff7900;font-weight:700;}#yiv3297290559 #yiv3297290559ygrp-reco {margin-bottom:20px;padding:0px;}#yiv3297290559 #yiv3297290559ygrp-sponsor #yiv3297290559ov li a {font-size:130%;text-decoration:none;}#yiv3297290559 #yiv3297290559ygrp-sponsor #yiv3297290559ov li {font-size:77%;list-style-type:square;padding:6px 0;}#yiv3297290559 #yiv3297290559ygrp-sponsor #yiv3297290559ov ul {margin:0;padding:0 0 0 8px;}#yiv3297290559 #yiv3297290559ygrp-text {font-family:Georgia;}#yiv3297290559 #yiv3297290559ygrp-text p {margin:0 0 1em 0;}#yiv3297290559 #yiv3297290559ygrp-text tt {font-size:120%;}#yiv3297290559 #yiv3297290559ygrp-vital ul li:last-child {border-right:none !important;}#yiv3297290559

Dwayne Verhey <yg@...>
 

Pete and David,-

Thanks for the advice. There's no sign of bulging, etc. on any of the
caps. Next time I'm in there I'll check for ripple. I believe one of the
big caps was replaced at some time in the past. It's neighbours are all
plain aluminium cans but there's one with a plastic sleeve over the can
and evidence of heat/flux around only those two solder joints on the board.

How old is old? this 475 dates from circa 1973. I habitually use flux
core silver bearing solder (62% tin, 36% lead, 2% silver) for
electronics. Is this suitable? (I've always used lead-free is for
potable water, and tin/lead is for drain pipes. ;-) )

I have one of those spring-loaded solder suckers (although the tip has
gotten too close to the heat a few times over the years). I also have
some braid. For heat, I have a 70 watt soldering station (Hakko FX888)
with a selection of tip shapes and generally set it a whisker over 350°
for that solder. Any red flags or advice there?

On 11/01/2016 10:39 PM, Peter Hildebrandt petertech99h@... [TekScopes] wrote:

Dwayne,
I'd go 50/50, inspect caps for corrosion, bulging cases, etc and start
sourcing replacement parts as you go and set them aside so they'll be
there when you need them. Some caps on double sided PCBs have plate
thru holes where the solder is hard to suck out without damage to the
board, so upgrade your solder iron and solder removal tools. Make sure
you have the right solder for your Tek scope, the very old vintage
scopes have silver bearing solder! If there's a lead-free solder
requirement, practice since this style solder is not so nice to work
with as 63/37 lead solders.

In my case, I check the AC ripple voltage across the caps on my
scope's linear power supplies, when it gets higher then it should I
know a cap is starting to fail.On 'newer' Tek scopes with switch-mode
power supplies you can't really do that though.
Have fun!
Pete


On Friday, January 1, 2016 10:08 PM, "David @DWH
[TekScopes]" <TekScopes@...> wrote:


For aluminum electrolytic capacitors, I leave them alone until one or
more fails or there are symptoms indicating that they are worn and
then I change all of them starting with the output capacitors for
switching power supplies and the input capacitors for linear power
supplies.

If they are old, I may change them during the process of repairing
something else while I have access anyway.

Solid tantalum capacitors are much the same except that they do not
have a real wear out mechanism and seem to fail destructively at which
point I will change any of the same type which were not properly
voltage derated.

On Fri, 01 Jan 2016 21:27:58 -0500, you wrote:

I joined this group for advice and insight for a newly-acquired 475 of
unknown provenance, and with your help, I found most of the problems
were with me, not the scope; only one minor repair, cleaning and some
tweaking were actually required to get it up and running.

As I lurk, about 99% of the repair advice here goes straight over my
head but I note a couple of recurring themes. Tektronix scopes may
contain custom, irreplaceable, components and failure of one component
may take out other bits and pieces.

As with other old electronics, capacitors seem to be mentioned a lot in
that second category. Particularly electrolytic and tantalum types. It's
almost automatic to replace original electrolytic capacitors in most
other old electronics, but I've seen two schools of thought with regards
to Tek scopes. A recent comment here suggested "don't fix it if it ain't
broke", and I have read that Tek used top quality components coupled
with circuit designs that did not stress the components, so failures are
supposedly the exception rather than the rule. And there's always the
risk of a ham-fisted amateur (ie: me) damaging things like traces when
replacing components. But then, there's a lot of people talking about
dry electrolytics and the smell of tantalum around here...

So what are your thoughts: would I be better off attempting some
preventative maintenance like replacing all the electrolytics or leaving
well enough alone until/unless something actually breaks?

<cross fingers and hope this doesn't hijack an existing thread>

Dwayne

 

On Mon, 11 Jan 2016 23:41:22 -0500, you wrote:

How old is old? this 475 dates from circa 1973.
My notes indicate that the 475 was offered from 1973 to 1983 so you
have one of the older ones but they are all what I would consider old.
Even under ideal conditions, aluminum electrolytic capacitors can "dry
out" after 20 years and no 475 is going to be younger than that.

You should be able to get an accurate production date by reading the
date codes on any semiconductors and sometimes passive parts in the
oscilloscope.

I habitually use flux
core silver bearing solder (62% tin, 36% lead, 2% silver) for
electronics. Is this suitable? (I've always used lead-free is for
potable water, and tin/lead is for drain pipes. ;-) )
I have a couple of spools of this type of solder although I prefer
eutectic Sn63Pb37 for routine work. The 2% silver solder is much
stronger and probably better in general although to me it seems to
oxidize faster. When soldering, I can smell the difference between
them.

Some surface mount parts use silver as a barrier metal on their
terminations (silver over palladium?) so 2% silver is needed to
prevent dissolving too much of the silver plating. Old Tektronix tube
era oscilloscopes with the ceramic strips need silver solder as well
for the same reason; I think Tektronix used 4% silver solder for them
but 2% silver solder should be good enough.

I have one of those spring-loaded solder suckers (although the tip has
gotten too close to the heat a few times over the years). I also have
some braid. For heat, I have a 70 watt soldering station (Hakko FX888)
with a selection of tip shapes and generally set it a whisker over 350°
for that solder. Any red flags or advice there?
The wattage of the soldering iron is not as important as being
temperature controlled. Your Hakko FX888 is likely better than
anything I use.

I just noticed Hakko's line of soldering stations which flood the area
around the tip with heated nitrogen to prevent oxidation. I wonder
how well that works.

With non-lead free solders, I try to use a lower temperature of 600F
to prevent lifting of the traces when doing rework on printed circuit
boards. For assembly and wired connections, I use 700F. With lead
free solders, the temperature needs to be higher which is hard on the
printed circuit boards so when I do rework, I melt some Sn63Pb37 into
the solder before desoldering to lower the melting point and then use
Sn63Pb37 for the rework.

For rework, fully activated rosin flux can make things much easier. I
like Kester 1544 but it is hard to find in small bottles.

Dwayne Verhey <yg@...>
 

On 13/01/2016 2:43 PM, David @DWH [TekScopes] wrote:

On Mon, 11 Jan 2016 23:41:22 -0500, you wrote:

How old is old? this 475 dates from circa 1973.
My notes indicate that the 475 was offered from 1973 to 1983 so you
have one of the older ones but they are all what I would consider old.
Even under ideal conditions, aluminum electrolytic capacitors can "dry
out" after 20 years and no 475 is going to be younger than that.
Ok, given that we have now established that my caps are probably twice
20 years, do you still advise waiting until problems actually appear,
rather than preemptive action?


I just noticed Hakko's line of soldering stations which flood the area
around the tip with heated nitrogen to prevent oxidation. I wonder
how well that works.
My Hakko is not that fancy. Mine is their low-end hobbyist station. But
recommended to me as better than what one might find in a hardware or
big-box store. As I understand it, temperature control is good but
wattage makes it possible. IE: sufficient wattage for the mass being
heated is necessary to maintain the set temperature in actual use.

With non-lead free solders, I try to use a lower temperature of 600F
to prevent lifting of the traces when doing rework on printed circuit
boards. For assembly and wired connections, I use 700F. With lead
free solders, the temperature needs to be higher which is hard on the
printed circuit boards so when I do rework, I melt some Sn63Pb37 into
the solder before desoldering to lower the melting point and then use
Sn63Pb37 for the rework.
Ok, °C vs °F. 700°F works out to 371°C and 600°F is about 315°C. Given
the silver raises the melting point a bit, I'm probably in the right
ball park for new, but perhaps a bit high for re-work. Good to know.
Thanks.

Paul Amaranth
 

Check the ripple value on the power supply test points. If it's in spec,
leave them be. If it's out of spec, then consider changing them out.
The actual values are in the service manual, but I think most of them
call for 5mV or less.

If you don't have another scope to check them, use a good dmm on an AC
range.

It's relatively easy to lift traces around those caps when removing them.
Also, Tek used the can as a jumper so you need to jumper the ground pins
together. A very nice repair can be done with a small PCB that accepts
a standard snap cap and uses the pinout of the old cap. There's gerber
files for that here:

http://www.eevblog.com/forum/testgear/tektronix-76037613-repairs/

I've used that type of adapter before and it makes for a very nice repair.

Use 105C caps if you can when replacing them.

Paul

On Wed, Jan 13, 2016 at 05:09:19PM -0500, Dwayne Verhey yg@... [TekScopes] wrote:
On 13/01/2016 2:43 PM, David @DWH [TekScopes] wrote:

On Mon, 11 Jan 2016 23:41:22 -0500, you wrote:

How old is old? this 475 dates from circa 1973.
My notes indicate that the 475 was offered from 1973 to 1983 so you
have one of the older ones but they are all what I would consider old.
Even under ideal conditions, aluminum electrolytic capacitors can "dry
out" after 20 years and no 475 is going to be younger than that.
Ok, given that we have now established that my caps are probably twice
20 years, do you still advise waiting until problems actually appear,
rather than preemptive action?


I just noticed Hakko's line of soldering stations which flood the area
around the tip with heated nitrogen to prevent oxidation. I wonder
how well that works.
My Hakko is not that fancy. Mine is their low-end hobbyist station. But
recommended to me as better than what one might find in a hardware or
big-box store. As I understand it, temperature control is good but
wattage makes it possible. IE: sufficient wattage for the mass being
heated is necessary to maintain the set temperature in actual use.

With non-lead free solders, I try to use a lower temperature of 600F
to prevent lifting of the traces when doing rework on printed circuit
boards. For assembly and wired connections, I use 700F. With lead
free solders, the temperature needs to be higher which is hard on the
printed circuit boards so when I do rework, I melt some Sn63Pb37 into
the solder before desoldering to lower the melting point and then use
Sn63Pb37 for the rework.
Ok, �C vs �F. 700�F works out to 371�C and 600�F is about 315�C. Given
the silver raises the melting point a bit, I'm probably in the right
ball park for new, but perhaps a bit high for re-work. Good to know.
Thanks.


!DSPAM:5696ca5e77571386779520!
--
Paul Amaranth, GCIH | Rochester MI, USA
Aurora Group, Inc. | Security, Systems & Software
paul@... | Unix & Windows

 

On Wed, 13 Jan 2016 17:09:19 -0500, you wrote:

On 13/01/2016 2:43 PM, David @DWH [TekScopes] wrote:

On Mon, 11 Jan 2016 23:41:22 -0500, you wrote:

How old is old? this 475 dates from circa 1973.
My notes indicate that the 475 was offered from 1973 to 1983 so you
have one of the older ones but they are all what I would consider old.
Even under ideal conditions, aluminum electrolytic capacitors can "dry
out" after 20 years and no 475 is going to be younger than that.
Ok, given that we have now established that my caps are probably twice
20 years, do you still advise waiting until problems actually appear,
rather than preemptive action?
I agree with Paul in this case. The 475 uses a linear power supply
and nothing is at risk as the aluminum electrolytic capacitors wear
out but once it becomes apparent that there is a problem, I would
change all of the input and output capacitors at the same time.

I just noticed Hakko's line of soldering stations which flood the area
around the tip with heated nitrogen to prevent oxidation. I wonder
how well that works.
My Hakko is not that fancy. Mine is their low-end hobbyist station. But
recommended to me as better than what one might find in a hardware or
big-box store. As I understand it, temperature control is good but
wattage makes it possible. IE: sufficient wattage for the mass being
heated is necessary to maintain the set temperature in actual use.
The relatively new Weller WES51 and WESD51 stations look like their
response to the inexpensive Hakko FX888 series and all of these would
have been considered high end not many years ago. What I would prefer
is a station like these but with an optional vacuum desoldering head
but I do no think that is an option.

I like that Hakko has all of the parts like elements and irons
available.

With non-lead free solders, I try to use a lower temperature of 600F
to prevent lifting of the traces when doing rework on printed circuit
boards. For assembly and wired connections, I use 700F. With lead
free solders, the temperature needs to be higher which is hard on the
printed circuit boards so when I do rework, I melt some Sn63Pb37 into
the solder before desoldering to lower the melting point and then use
Sn63Pb37 for the rework.
Ok, °C vs °F. 700°F works out to 371°C and 600°F is about 315°C. Given
the silver raises the melting point a bit, I'm probably in the right
ball park for new, but perhaps a bit high for re-work. Good to know.
Thanks.
It helps a lot to have an iron with large heat capacity and high power
like your Hakko. It makes working at lower temperatures much easier.

John Clark
 

Paul, thanks for this tip! That's the first install of modern snap caps I've seen that I thought looked presentable and reliable.

My own 475 will likely need the filter caps replaced at some point but I was talked out of doing it right now since ripple and noise all checks out good and the scope is working flawlessly. I've never ordered custom PCB's before. The Gerber files are linked there, and I've downloaded them, but is it as easy as sending those files to someone and they just make them for you? The guy said iTead made them for about $1 each but I wouldn't know how to go about doing that on their website.

A very nice repair can be done with a small PCB that accepts
a standard snap cap and uses the pinout of the old cap. There's gerber

files for that here:



http://www.eevblog.com/forum/testgear/tektronix-76037613-repairs/



I've used that type of adapter before and it makes for a very nice repair.



Use 105C caps if you can when replacing them.



Paul

Dwayne Verhey <yg@...>
 

Paul,-

Ok, I found the ripple spec in the service manual:

"g. Check ripple on supplies (except +110 V supply) is
2 mV p-p maximum; +110 V supply ripple is 1V p-p
maximum."

I do have another scope (although it's a Rigol -- am I allowed to use an entry level Chinese scope to troubleshoot a Tek? <lol>) so I'll check the ripple at the various test points with that.

Thanks.

On 13/01/2016 5:20 PM, Paul Amaranth paul@... [TekScopes] wrote:

Check the ripple value on the power supply test points. If it's in spec,
leave them be. If it's out of spec, then consider changing them out.
The actual values are in the service manual, but I think most of them
call for 5mV or less.

Dwayne Verhey <yg@...>
 

Paul and David,-

How much does frequency of the ripple matter?

Spec is 2mv ripple max at the test points (except 1v @ 110v test point). If I limit the BW to 20MHz (Tek) or filter above 100Hz (Rigol) then all test points are in spec. But there's some high-frequency stuff, looks like she's picking up some RF, primarily around 110MHz. Typical HF ripple across all test points is

Rigol (measurement function): 5.6mV P-P (of the entire noise envelope) or 1.9mV (x1 probe)

Tek 475 measuring itself: 5.8mV P-P (again, of the entire envelope) (x1 probe)

Bench meter: 0.0033 +/-0.0002 VAC RMS (1000Mohm input impedance)

Interestingly, at the Rigol's full bandwidth and using the manual cursors, I can pick out the biggest contributor to the HF noise envelope: a 400nS pulse recurring at a frequency of 60Hz.

On 13/01/2016 5:20 PM, Paul Amaranth paul@... [TekScopes] wrote:

Check the ripple value on the power supply test points. If it's in spec,
leave them be. If it's out of spec, then consider changing them out.
The actual values are in the service manual, but I think most of them
call for 5mV or less.

If you don't have another scope to check them, use a good dmm on an AC
range.

Tothwolf
 

On Wed, 13 Jan 2016, John Clark johnclark05@... [TekScopes] wrote:

Paul, thanks for this tip! That's the first install of modern snap caps I've seen that I thought looked presentable and reliable.

My own 475 will likely need the filter caps replaced at some point but I was talked out of doing it right now since ripple and noise all checks out good and the scope is working flawlessly. I've never ordered custom PCB's before. The Gerber files are linked there, and I've downloaded them, but is it as easy as sending those files to someone and they just make them for you? The guy said iTead made them for about $1 each but I wouldn't know how to go about doing that on their website.
Here is what I did with some of my scopes:

http://strudel.ignorelist.com/~tothwolf/photos/Tektronix_2213/Tektronix_2213_2/1600/IMG_9085.1600.jpg
http://strudel.ignorelist.com/~tothwolf/photos/Tektronix_2213/Tektronix_2213_2/1600/IMG_8873.1600.jpg

I was originally going to use radial mount parts which had a 12mm lead spacing, but I discovered a 10mm snap-in fit better so I used those instead. I used two small sections of tinned copper insulated with teflon tubing to jumper the ground connections between the 3 outer pads. I forget now which AWG I used but I think it was either 18 AWG or 16 AWG.

I found this little factory "oops" in scope #3 while replacing the safety capacitor located under the IEC inlet filter:

http://strudel.ignorelist.com/~tothwolf/photos/Tektronix_2213/Tektronix_2213_3/1600/IMG_9483.1600.jpg

I also fixed this problem in scope #2. The long leads of those film capacitors are not all at the same potential and two were shorted together (note that the service manual has the board and schematic designations for these mixed up). I insulated them with teflon tubing in all 3 scopes to keep it from ever becoming an issue again.

http://strudel.ignorelist.com/~tothwolf/photos/Tektronix_2213/Tektronix_2213_2/1600/IMG_8989.1600.jpg
http://strudel.ignorelist.com/~tothwolf/photos/Tektronix_2213/Tektronix_2213_2/1600/IMG_9043.1600.jpg

More photos can be found here:

http://strudel.ignorelist.com/~tothwolf/photos/Tektronix_2213/

There are more things I still plan to do on these scopes but I haven't had a chance to get back to them. I also never finished photographing scope #3.

Siggi
 

On Fri, 15 Jan 2016 at 10:58 Dwayne Verhey yg@... [TekScopes] <
TekScopes@...> wrote:



Paul and David,-

How much does frequency of the ripple matter?

The noise and ripple is typically specified at 20MHz bandwidth. Also
beware that it's specified as seen on an analog scope of the time.
Digital scopes may show higher noise and ripple, as depending on settings
such as memory depth. Effectively they can be set to more persistence than
the phosphor of the old-style analog scope.

Dwayne Verhey <yg@...>
 

Siguour,-

I'm curious where you get the 20MHz specification. The manual calls for
a scope with "Bandwidth, DC to 100 mega-hertz; minimum deflection factor, five millivolts/division; accuracy, within 3%, and the example
given is the Tektronix 454A Oscilloscope with two P6054 probes (which
Google says are 150MHz and 250MHz, respectively).

I'm not saying you are wrong, the test scope is also to be used for
other things like the high-speed timing adjustment. I just don't see
anything in the manual about 20 MHz (mind you, these manuals often
assume some basic level of prior knowledge and experience, which I don't
have).

I certainly get your point about the persistence of a digital scope
compared to the old analogue. Even with the intensity cranked up to
maximum, the 475 trace is pretty faint at the fastest time base settings
using the delayed sweep function. Even though the 475 is rated at 200MHz
bandwidth and the Rigol is rated equivalent to 150MHz (mine is upgraded
to the DS1152E-EDU spec), It's a darn sight easier to catch and study a
transient or high-speed waveform displayed on the digital scope!

That said, at slower time base settings, I really like the "feel" of the
475 analogue display compared to the Rigol. The variations in intensity
and the smoothness of the curves highlight the pixellation of the
Rigol's display. Now if only it was in colour...


On 15/01/2016 8:45 PM, Sigurður Ásgeirsson siggi@... [TekScopes]
wrote:


The noise and ripple is typically specified at 20MHz bandwidth. Also
beware that it's specified as seen on an analog scope of the time.
Digital scopes may show higher noise and ripple, as depending on settings
such as memory depth. Effectively they can be set to more persistence than
the phosphor of the old-style analog scope.


Siggi
 

Hey Dwayne,

In the service manual for the 2400-series scopes it clearly specified to
use 20MHz bandwidth for PSU ripple measurement. I also believe it's a
de-facto standard for this, as noise is a monotonic function of bandwidth.
That being said I checked the service manual for my 485, which just lists
acceptable ripple without further spec.

Siggi
On Fri, Jan 15, 2016 at 21:52 Dwayne Verhey yg@... [TekScopes] <
TekScopes@...> wrote:



Siguour,-

I'm curious where you get the 20MHz specification. The manual calls for
a scope with "Bandwidth, DC to 100 mega-hertz; minimum deflection
factor, five millivolts/division; accuracy, within 3%, and the example
given is the Tektronix 454A Oscilloscope with two P6054 probes (which
Google says are 150MHz and 250MHz, respectively).

I'm not saying you are wrong, the test scope is also to be used for
other things like the high-speed timing adjustment. I just don't see
anything in the manual about 20 MHz (mind you, these manuals often
assume some basic level of prior knowledge and experience, which I don't
have).

I certainly get your point about the persistence of a digital scope
compared to the old analogue. Even with the intensity cranked up to
maximum, the 475 trace is pretty faint at the fastest time base settings
using the delayed sweep function. Even though the 475 is rated at 200MHz
bandwidth and the Rigol is rated equivalent to 150MHz (mine is upgraded
to the DS1152E-EDU spec), It's a darn sight easier to catch and study a
transient or high-speed waveform displayed on the digital scope!

That said, at slower time base settings, I really like the "feel" of the
475 analogue display compared to the Rigol. The variations in intensity
and the smoothness of the curves highlight the pixellation of the
Rigol's display. Now if only it was in colour...

On 15/01/2016 8:45 PM, Sigurður Ásgeirsson siggi@... [TekScopes]
wrote:




The noise and ripple is typically specified at 20MHz bandwidth. Also
beware that it's specified as seen on an analog scope of the time.
Digital scopes may show higher noise and ripple, as depending on settings
such as memory depth. Effectively they can be set to more persistence
than
the phosphor of the old-style analog scope.




[Non-text portions of this message have been removed]

Marian B
 

The average HP 3400A has a -3 dB bandwidth of about 5 Hz to 20...25 MHz.
I guess that in fact the 3400A set the 20 MHz bandwidth standard for
this kind of measurement. Maybe even Tek themselves used 3400A's?

(The 10 MHz spec of the 3400A is a +- 5 % bandwidth i.e. +- 0.5 dB)

Cheers, Marian

On 16.01.2016 04:37, Sigurður Ásgeirsson siggi@... [TekScopes] wrote:


Hey Dwayne,

In the service manual for the 2400-series scopes it clearly specified to
use 20MHz bandwidth for PSU ripple measurement. I also believe it's a
de-facto standard for this, as noise is a monotonic function of bandwidth.
That being said I checked the service manual for my 485, which just lists
acceptable ripple without further spec.

Siggi
On Fri, Jan 15, 2016 at 21:52 Dwayne Verhey yg@... [TekScopes] <
TekScopes@...> wrote:



Siguour,-

I'm curious where you get the 20MHz specification. The manual calls for
a scope with "Bandwidth, DC to 100 mega-hertz; minimum deflection
factor, five millivolts/division; accuracy, within 3%, and the example
given is the Tektronix 454A Oscilloscope with two P6054 probes (which
Google says are 150MHz and 250MHz, respectively).

I'm not saying you are wrong, the test scope is also to be used for
other things like the high-speed timing adjustment. I just don't see
anything in the manual about 20 MHz (mind you, these manuals often
assume some basic level of prior knowledge and experience, which I don't
have).

I certainly get your point about the persistence of a digital scope
compared to the old analogue. Even with the intensity cranked up to
maximum, the 475 trace is pretty faint at the fastest time base settings
using the delayed sweep function. Even though the 475 is rated at 200MHz
bandwidth and the Rigol is rated equivalent to 150MHz (mine is upgraded
to the DS1152E-EDU spec), It's a darn sight easier to catch and study a
transient or high-speed waveform displayed on the digital scope!

That said, at slower time base settings, I really like the "feel" of the
475 analogue display compared to the Rigol. The variations in intensity
and the smoothness of the curves highlight the pixellation of the
Rigol's display. Now if only it was in colour...

On 15/01/2016 8:45 PM, Sigurður Ásgeirsson siggi@... [TekScopes]
wrote:




The noise and ripple is typically specified at 20MHz bandwidth. Also
beware that it's specified as seen on an analog scope of the time.
Digital scopes may show higher noise and ripple, as depending on
settings
such as memory depth. Effectively they can be set to more persistence
than
the phosphor of the old-style analog scope.



Dwayne Verhey <yg@...>
 

Siggi,-

It sounds like the balance of probability is that at the time the 400
series scopes were introduced, 20MHz was a typical scope BW, so it
wasn't necessary to mention it; but by the time the 2400 series was
introduced there were enough 100 and 200MHz scopes around that Tek's
techs figured they better clarify things for people like me.

On 15/01/2016 10:37 PM, Sigurður Ásgeirsson siggi@... [TekScopes]
wrote:

Hey Dwayne,

In the service manual for the 2400-series scopes it clearly specified to
use 20MHz bandwidth for PSU ripple measurement. I also believe it's a
de-facto standard for this, as noise is a monotonic function of bandwidth.
That being said I checked the service manual for my 485, which just lists
acceptable ripple without further spec.

Siggi
On Fri, Jan 15, 2016 at 21:52 Dwayne Verhey yg@... [TekScopes] <
TekScopes@...> wrote:

 

The high frequency hash is external to the oscilloscope and just
represents RF in the ambient environment. Some may be conducted noise
via a ground loop between the 475 and Rigol through the test probe
ground connection.

In practice, measurements which use the probe's ground clip at high
vertical sensitivities are always limited by ambient RF noise. Either
the probe's ground clip needs to be discarded and a coaxial connection
made to the probe or the bandwidth needs to be limited. A coaxial
connection to the probe tip works great but is a hassle and a power
supply ripple measurement does not need that kind of bandwidth anyway.

The power supply ripple comes in three forms:

1. The 50/60 Hz line power is rectified to produce 100/120 Hz full
wave ripple which is attenuated by the input capacitor and line
response of the regulator. If the input capacitor lacks enough
capacitance to hold the input voltage above the regulator's dropout
voltage, then this ripple will become pronounced.

This ripple can be more easily measured with an oscilloscope by using
line triggering however an AC voltmeter works well also. If you use
line triggering with a DSO, then averaging can be used to remove all
noise except for the 50/60 Hz ripple and harmonics allowing very low
noise measurements to be made.

If you use an AC voltmeter, then an estimate can be done of peak to
peak ripple which should be about 3 times the average or RMS (they are
about the same in this case) but beware that most meters work poorly
at the lowest end of their AC ranges and they will include any low
frequency noise in their measurement. RMS reading meters can also
take considerable time to settle at the lowest end of their range so
this is an application where average reading AC would be better.

2. High frequency ripple is suppressed (decoupled) by the output
capacitor. If the output capacitor's ESR rises which happens as an
aluminum electrolytic dries out, then it will become more pronounced.
This measurement needs higher bandwidth but 5 to 20 MHz is plenty.

3. Medium frequency ripple depends on the regulator's load response
and somewhat on the output capacitor.

So if you find excessive 100/120 Hz ripple, then the input capacitor
is bad. If you find excessive high frequency ripple, then the output
capacitor is bad.

I like using my 2230 in peak detect mode for power supply ripple
measurements but peak to peak acquisitions are pessimistic and can
show more noise than is really there; they are good for a worse case
measurement. Perhaps more useful are a 7A13 (1 mV/div at 5 MHz), 2225
(500 uV/div at 5 MHz), or even a 7A22 which is much more sensitive but
limited to 1 MHz. Any analog oscilloscope though with a 20 MHz
bandwidth limit is practically as good though.

Note that while using a x1 probe automatically limits input bandwidth
to about 20 or 30 MHz, it is still worth using any switchable
bandwidth limit function because this will limit the noise coming from
the test oscilloscope itself which can be considerable at its most
sensitive setting.

I do not know what your 60 Hz 400nS noise spike is. If it was 120 Hz,
then I would suspect a phase controlled light dimmer or perhaps a
ballasted CFL or LED lamp. I would connect the probe ground and tip
both the 475 ground and see if it is still there and represents
external RFI or maybe a ground loop. Honestly I do not trust Rigol to
make an oscilloscope which does not suffer from a problem like that.

Your measurements seem reasonable to me so I would not change the
capacitors yet unless you have some other reason. The acid test of
course is proper operation of the 475 including vertial noise and
horizontal jitter.

On Fri, 15 Jan 2016 10:58:12 -0500, you wrote:

Paul and David,-

How much does frequency of the ripple matter?

Spec is 2mv ripple max at the test points (except 1v @ 110v test point).
If I limit the BW to 20MHz (Tek) or filter above 100Hz (Rigol) then all
test points are in spec. But there's some high-frequency stuff, looks
like she's picking up some RF, primarily around 110MHz. Typical HF
ripple across all test points is

Rigol (measurement function): 5.6mV P-P (of the entire noise envelope)
or 1.9mV (x1 probe)

Tek 475 measuring itself: 5.8mV P-P (again, of the entire envelope) (x1
probe)

Bench meter: 0.0033 +/-0.0002 VAC RMS (1000Mohm input impedance)

Interestingly, at the Rigol's full bandwidth and using the manual
cursors, I can pick out the biggest contributor to the HF noise
envelope: a 400nS pulse recurring at a frequency of 60Hz.

On 13/01/2016 5:20 PM, Paul Amaranth paul@... [TekScopes] wrote:

Check the ripple value on the power supply test points. If it's in spec,
leave them be. If it's out of spec, then consider changing them out.
The actual values are in the service manual, but I think most of them
call for 5mV or less.

If you don't have another scope to check them, use a good dmm on an AC
range.

Dwayne Verhey <yg@...>
 

David,-

Thanks for taking the time to provide such an easy to understand yet detailed explanation. I am slowly trying to disprove the theory that "you can't teach an old dog new tricks" and posts like that go a long way in the endeavour!

On 20/01/2016 1:54 AM, David @DWH [TekScopes] wrote:

The high frequency hash is external to the oscilloscope and just
represents RF in the ambient environment. Some may be conducted noise
via a ground loop between the 475 and Rigol through the test probe
ground connection.
(snip)

 

That post sort of got away from me. It just kept going like the
Energizer bunny.

On Wed, 20 Jan 2016 09:55:44 -0500, you wrote:

David,-

Thanks for taking the time to provide such an easy to understand yet
detailed explanation. I am slowly trying to disprove the theory that
"you can't teach an old dog new tricks" and posts like that go a long
way in the endeavour!

On 20/01/2016 1:54 AM, David @DWH [TekScopes] wrote:

The high frequency hash is external to the oscilloscope and just
represents RF in the ambient environment. Some may be conducted noise
via a ground loop between the 475 and Rigol through the test probe
ground connection.
(snip)